AI News

The Sovereign AI fund is providing funding and compute capacity to Potential Sciences.

The next phase of the AI investment boom is taking shape, and investors are backing companies that can generate AI content, protect corporate data, and build the infrastructure needed to make advanced technology commercially viable. Leading today's startup funding activity is China's Sand.ai, which secured $200 million to develop generative video models. Cybersecurity startup Sentra

AI Olympics: Il 3 dicembre a Milano Alterna e Microsoft trasformano l’AI agentica in una gara a squadre. Il pubblico assegnerà le medaglie ai vincitori

Meta's AI agent can now take calls as a customizable animated avatar, ahead of the launch of a Tamagotchi-like wearable designed to bring Muse to life.

Generate, edit, subtitle, dub, and translate videos using the Kapwing MCP in Claude or ChatGPT.

Build and deploy enterprise agents, workflows and apps faster with Mistral Studio and Le Chonk full control over your models, data and infrastructure.

Trained on 300 hours of generated motion in simulated rooms, the onboard system completed 91% of 360 real-world trials without any teleoperation data.

SpaceX's bold claim that it's capable of building 10 GW of compute in 2027 may be plausible — analysts dig through the breakneck speed of Colossus buildouts as evidence

Make intelligence open and accessible to all

ERNIE 5.5 Flash Base: a 280B-A14B hybrid-attention MoE pretrained on 33.5T tokens. An open, efficient pretraining base.

Language agents increasingly use persistent memory to record user preferences and procedures across sessions. We show that memory from benign interactions can cause unsafe or unreliable behavior on later tasks. We call this failure the memory trap. To study this issue, we compare behavior in sessions with and without memory across six frontier models (including Claude Sonnet, Opus, and Fable models and GPT-5.6). We evaluate these models in a range of settings. First, in autonomy experiments, we find that requests to skip confirmation during low-stakes coding chores lead agents to skip confirmation and verification on simulated high-stakes tasks in both coding and everyday life. We next study memories formed when users seek emotional support, or describe harm from uncritical feedback, and find these memories bias later judgments of mathematical work, arguments, and poems toward sycophancy or over-critique (respectively). Memory notes written during legitimate stale-test repairs lead agents to weaken correct tests that catch broken safeguards; these memories also increase cheating on ImpossibleBench, with larger increases for stronger Sonnet and Opus models. We show that memory traps can arise both from overgeneralizations in memory writing, and when acting on memory. These findings demonstrate that even benign interactions can lead to serious risks from memory systems, and thus motivate the need for new research into ensuring that agents preserve the scope of their memories.

Can large language models (LLMs) advance theoretical computer science and resolve its long-standing open conjectures? To study progress toward this goal, we introduce TCSExam, a unified framework and benchmark covering the full range of TCS research. It evaluates theorem proving, solver engineering, and constructions alongside direct attempts to resolve open conjectures, capturing both concrete research accomplishments and progress toward long-standing research goals. TCSExam-Full contains 100 expert-reviewed tasks, with submissions checked by the Lean kernel or an executable task checker. TCSExam-Lite selects 20 tasks, five per track, reducing the number of task attempts by 80% at the same per-attempt budget. We evaluate eight agents on Full and nine on Lite. The TCSExam Index combines all four tracks: GPT-6 Astra leads Full (71.9), and Opus 5.5 leads Lite (72.3). Full and Lite rankings are strongly aligned across eight shared agents (Spearman's $ρ=0.93$), supporting Lite as a compact, discriminative benchmark. The evaluations yield improved solver outputs and mathematical constructions, together with independently verified special cases and bounds from open-conjecture attempts. TCSExam provides an extensible framework with task sources, Docker build recipes, and evaluation tools for reproducible use.

AI-Scientists arrived, but a world for them to grow in did not. Human scientists take theirs for granted $-$ centuries of open problems and shared ideas, waiting in the scientific literature. To AI-Scientists, that world is readable but not runnable, so we build Scientist World for them: every research paper becomes an executable environment. Our autonomous pipeline extracts each paper's open-ended research question and reproduces its proposed method as the baseline; an AI-Scientist entering the environment is evaluated against what the human authors achieved. We release the first 1,000 verified environments. With the world as an evaluation, we found that frontier AI-Scientists can already surpass the human researchers 74% of the time. As a record of how AI does science, it contains 6,400 hours of research traces, with every idea, experiment, failure, and success. As a training ground, we are able to raise (+158%) a small model to match frontier models in research ideation. At last, AI-Scientists' own world has arrived, grown from and growing with science itself.

Kalepa raised $75 million in a Series B at a $470 million valuation, led by Updata Partners. The New York-based insurance AI start-up previously closed a

The deal is expected to support the development of DBX Agent Cloud, a new storage platform aimed at developers building AI models and agents.

Il padiglione di ITA-ICE ospita startup e PMI all'edizione 2026 della manifestazione diventata una delle principali fiere tech europee

Apple Logic Pro 13 is out now, featuring a redesigned DAW interface, a new smart browser, new ways to use it, and AI features not for all

Ask Claude about your money and get straight answers from the accounts you connect.

Pixel combines a low-cost two-armed robot with human demonstration tools and whole-body learning. Its companion Pixel-UMI study reports 87% average task progress across five tasks.

Companies tied to the AI buildout are all selling off as concerns mount about a slowdown following the midterm elections.

Google is investing to build its first Wyoming data center, creating thousands of jobs and innovative carbon capture. Get the details.

Today we are introducing the second generation of Inkling models with the release of Inkling 2 Small.

A Blog post by NVIDIA on Hugging Face

Test-time scaling methods spend additional compute at deployment to improve a large language model's performance on a hard target problem. Most methods rely on a verifier that scores candidate solutions for correctness, using it to select among or refine the model's outputs. In many deployment regimes, however, a strong verifier on the target problem is unavailable. We introduce Test-Time Self-Play (TTSP), an approach to test-time scaling using asymmetric self-play without reliance on a verifier. On a difficulty-calibrated subset of LiveCodeBench v6, TTSP is the best-performing method at every verifier strength we evaluate, and the gap to the strongest baseline widens as the verifier weakens; in the no-verifier limit, TTSP yields a 35\% absolute improvement in mean@8. The gains hold across a range of verifier strengths, except when the base policy has near-zero coverage of the correct answer.

ML code should run where researchers choose, not where it was first written. Today, machine learning code mainly lives in two ecosystems, PyTorch on NVIDIA GPUs and JAX on Google TPUs. Migrating repositories between them means building and testing across frameworks and hardware. This migration is (1) a practical need, since it unlocks different hardware for researchers, and (2) a natural benchmark: it involves different frameworks and hardware, covers whole repositories, and is long-horizon and challenging. We introduce XStack, a benchmark that evaluates agents on migrating ML repositories between these two ecosystems. Each agent receives a PyTorch repository and a clean TPU machine and must deliver a JAX repository. Its output passes if it meets our grading criteria, based on the source repository's published result, authenticated and calibrated by us. Migration by agents remains unreliable: Claude Opus 5 passes 75%, GPT-5.6-sol 45%, and every other agent at most 40%. XStack is a step toward freeing ML code from its ecosystem. We release XStack to make progress toward reliable migration across frameworks and hardware measurable.